Meta launches desktop AI as South Korea invests $502.8 million

Key Takeaways

• TrojAI and Harness introduced new AI security tools, including agent-led red teaming, runtime intelligence, real-time protection for coding assistants, and an AI Firewall for DevSecOps.
• A Robert Half study found 67% of hiring managers believe AI-generated resumes negatively impact hiring due to generic content and lack of authenticity.
• AI in hiring, including Applicant Tracking Systems and video interviews, can create barriers for job seekers and potentially perpetuate biases.
• Meta's AI startup Manus launched a desktop application, allowing its AI agent to run locally on personal devices and access files, similar to OpenClaw.
• The conflict in Iran has disrupted the global helium supply, a critical coolant for AI chips, potentially leading to shortages and supply chain delays.
• South Korea plans to invest approximately US$502.8 million (750 billion won) over two years to commercialize 246 AI-enabled products and services.
• Southern Connecticut State University is expanding AI and quantum technology training for manufacturing professionals, responding to a 40% rise in AI skill job postings in Connecticut.
• Advantech is showcasing edge AI innovations with NVIDIA Jetson Thor for applications in robotics, medical, logistics, and retail.
• The rise of AI companions like Replika and Character.AI indicates a societal trend towards on-demand, low-effort, and personalized friendships.
• A man was arrested for using an AI-edited photo, bearing a Meta AI watermark, to falsely claim car damage, highlighting potential for AI misuse in fraud.

TrojAI boosts AI security with new agent testing and protection tools

TrojAI has released new features to secure enterprise AI systems, going beyond basic prompts. These include Agent-Led AI Red Teaming, which uses autonomous agents to test AI models and applications against security frameworks like OWASP and MITRE. Additionally, Agent Runtime Intelligence offers deep visibility into how AI agents behave during operation. The company also introduced Real-Time Protection of Coding Agents to safeguard AI coding assistants from risks like exposed secrets and prompt injection attacks.

Harness launches AI security and secure coding tools for DevSecOps

Harness has introduced AI Security and Secure AI Coding tools to enhance its DevSecOps platform for the AI era. AI Security helps discover, test, and protect AI in applications, including an AI Firewall to block attacks and inspect inputs/outputs. Secure AI Coding integrates with AI coding assistants to scan code as it's written, identifying vulnerabilities in real-time. These tools aim to cover the entire AI lifecycle, from code generation to production.

TrojAI secures AI agents beyond prompts with new features

TrojAI has enhanced its platform to secure agentic AI, moving beyond just prompt-level protection. New features include Agent-Led AI Red Teaming for advanced testing of AI agents and models, mapping results to industry standards like OWASP and MITRE. Agent Runtime Intelligence provides detailed insights into AI agent behavior during operation. The company also offers Real-Time Protection for Coding Agents to secure AI assistants used in software development.

AI resumes hurting job applications, 67% of managers say

A study by Robert Half found that 67% of hiring managers believe AI-generated resumes are negatively impacting the hiring process. Key issues include a lack of authenticity, generic content, and over-reliance on AI without proper customization or proofreading. To stand out, job seekers are advised to personalize their applications, quantify achievements, highlight soft skills, tailor resumes to each job, and proofread meticulously. Networking also remains crucial for getting noticed.

AI in hiring creates barriers for job seekers, study finds

Millions of job seekers face challenges due to AI in the hiring process, with Applicant Tracking Systems (ATS) scanning and ranking resumes. Critics argue these AI systems can perpetuate biases and screen out qualified candidates if resumes don't match specific keywords. AI is also used in video interviews, analyzing expressions and tone, but lacks transparency and can be biased. While AI aims for efficiency, experts suggest making it work better for both employers and applicants.

Seth MacFarlane discusses AI's Hollywood impact and show futures

Seth MacFarlane spoke with Yahoo Entertainment about artificial intelligence's growing role in Hollywood, including its potential impact on filmmaking and TV production. He also addressed the future of his long-running animated series 'Family Guy,' indicating no immediate end is planned. MacFarlane also shared updates on the upcoming second season of his live-action 'Ted' series, hinting at new storylines.

Meta's Manus AI agent now runs on desktops, competes with OpenClaw

Meta's AI startup Manus has launched a desktop application allowing its AI agent to run directly on users' personal devices via a 'My Computer' feature. Previously cloud-based, the agent can now access local files and applications, similar to the open-sourced OpenClaw. While Manus is a paid service, it offers features like file management and app creation. Manus emphasizes user control over task execution to address potential security concerns.

Iran war disrupts helium supply vital for AI chips

The conflict in Iran has disrupted the global helium supply, impacting artificial intelligence chip makers and other industries. The closure of the Strait of Hormuz has trapped a significant portion of the world's commercial helium, a crucial coolant for advanced microchips and MRI scanners. While a surplus currently exists, experts predict a shortage, with potential price increases and supply chain delays affecting chip fabrication, especially in Asia.

Southern Connecticut State to train workers in AI and quantum tech

Southern Connecticut State University (SCSU) is expanding its workforce training programs in artificial intelligence and quantum technologies to meet growing demand. The initiative aims to prepare current and future manufacturing professionals to use AI tools for tasks like quality control and production efficiency. A new course, Quantum in Manufacturing 101, will also introduce quantum computing concepts. Connecticut saw a 40% increase in job postings requiring AI skills.

South Korea invests $500M to commercialize AI products

South Korea plans to invest approximately 750 billion won (US$502.8 million) over the next two years to commercialize artificial intelligence-enabled products. This initiative is part of a larger plan to build a 'super-innovation economy' and accelerate AI-driven economic transformation. The government will support the development and launch of 246 AI products and services, focusing on those that can reach the market within one to two years.

Advantech showcases edge AI with NVIDIA Jetson Thor

Advantech is demonstrating its edge AI and physical AI innovations at NVIDIA GTC 2026, utilizing technologies like NVIDIA Jetson Thor. The company is showcasing platforms for robotics, medical devices, logistics, and retail, designed for performance and scalability at the edge. Advantech's solutions aim to help enterprises deploy AI efficiently in real-world applications, integrating hardware, software frameworks, and industry partnerships.

AI companions reveal trends in human friendship

The rise of AI companions like Replika and Character.AI shows a shift towards friendships that are on-demand, low-effort, and personalized. Sociologists suggest these AI 'friends' reflect current trends in human connection, influenced by years of mediated communication through screens. While AI companions may not replace human friendships entirely, they highlight a growing preference for convenient, easily accessible connections, driven by increasing isolation and a culture of hyper-individualism.

Man arrested for faking car damage claim with AI photo

A Central New York man, Brent W. Hubbard, was arrested for allegedly trying to scam police with an AI-edited photo. He claimed his car was damaged by ice from a police vehicle and submitted a doctored image to support his reimbursement request. Cicero Police Chief Steve Rotunno stated the photo had a Meta AI watermark and that damage to the car could not be found. Hubbard has admitted to falsifying the incident, leading to charges of falsely reporting an incident.

Sources

• TrojAI Extends Enterprise AI Security with Agent-Led Red Teaming, Runtime Intelligence, and Coding Agent Protection
• Harness Introduces AI Security and Secure AI Coding Tools
• TrojAI unveils new capabilities to secure agentic AI beyond the prompt layer
• AI Resumes Are Sabotaging The Hiring Process, 67% Of Managers Reveal
• Computer says no. Are AI interviews making it harder to get a job?
• Seth MacFarlane on AI in Hollywood, How Long 'Family Guy' Will Continue, and 'Ted' Season 2
• Meta's Manus launches desktop app to bring its AI agent onto personal devices amid OpenClaw craze
• The Iran war disrupts global helium supply and artificial intelligence chip makers
• Southern Connecticut State to Offer Training in AI, Quantum Tech
• S. Korea to invest 750 bil. won in next 2 yrs to commercialize AI-enabled products
• Advantech shows robotics, medical AI, and industrial edge products using NVIDIA Jetson Thor
• What AI ‘Friends’ Reveal About Human Friendship
• Central NY man accused of trying to scam police with AI-edited photo. They were on to him quickly.